NYC Panel Session with Hidden Forces

Thursday, October 19, 2017

Listen on iTunes
Watch on YouTube
Visit HiddenForcesPod.com

Demetri: Thank you for everyone coming tonight. I really appreciate it. We have a real packed house. I want to thank The Assemblage for making the space available on such short notice, as well as the panelists for being here.

This panel came together because I had the opportunity to meet somebody by the name of Dr. Leemon Baird, who is the inventor of Hashgraph – a completely new distributed ledger technology – not blockchain. I had Leemon on the program and I was really blown away by this technology, and I wanted to learn more about it.

There is a lot of hype in the blockchain and crypto space. I thought this would be a great opportunity to bring in something totally new, and also open it up to the audience to ask questions about what this technology is, how it’s different from blockchain, and how the protocol works.

Why don’t we start, though, with the panel first introducing themselves?

Mance: Hello, my name is Mance Harmon. I am the co-founder with Leemon of Swirlds, Inc., which is the company that is commercializing the Hashgraph consensus algorithm. A little bit about my background – I am a computer scientist, with a couple of degrees, most recently from UMass. Leemon and I actually first started working together back at UMass in 1993. We worked for the Air Force as senior scientists for machine intelligence. Our formal backgrounds are in machine learning. We also both taught computer science at the Air Force Academy. Leemon was a full professor there. I was head of cyber security as a course director.

I managed a large software program for the missile defense agency, and was told I have to share this – how many of you remember the movie, War Games from the 1980’s? So, I managed the team – 160 people who built the real thing. That was a lot of fun, and of course, it’s now being used as the simulator for figuring out how to play these ballistic missile defense war games – essentially the strategy for protecting the U.S. and our allies against missiles.

Leemon and I started a company in the space of identity. We sold that company to a Fortune 500. I was the senior executive for product security for that same Fortune 500. We then started another company, also in identity. We sold that to private equity. After that, I worked for an identity player as the head of labs, and was their senior executive on staff for architecture.

Then, I got involved in this space. We started looking at blockchain for identity. Leemon, for the past five years was trying to solve the problem [distributed consensus], which he was ultimately able to do. We call it Hashgraph. And, that was quite a journey.

One other thing I should mention about Leemon is that he has a PhD from Carnegie Mellon in computer science, and has the notable distinction of setting the record for the history of the school in getting his PhD in the shortest amount of time – 2 years and 9 months – quite a feat.

What we’re going to talk about tonight is Leemon’s baby. We’re not from here – we’re from Texas. I’ve heard some people say, “How could something like this come from Texas?” And, that’s a fair question. But, we’re southerners from Texas and glad to be here talking about Hashgraph with the crypto community in New York City.

Jordan Fried: I am Jordan Fried, and I spent the last six years over in Budapest, Hungary of all places. I’m not originally a New Yorker, and am also new to the city. I sort of came in the crypto space accidently in that the previous software company that I ran was a virtual private network service. The number one demand we had from customers was to accept bitcoin as a payment method, and I said, “What the hell is bitcoin?” So, we integrated with BitPay and started receiving bitcoin. It wasn’t until the value went up, though, that I started to realize what this thing was.

That company was acquired at the beginning of this year, and then I basically begged my wife to move to New York City to figure out what to do next. That’s where both Andrew and Dr. Baird come in.

Dr. Baird is one of those guys you meet, and you know your life will never be the same after you meet him – he’s one of those brilliant minds. What he’s invented here and what we’re going to talk about is a testament to that.

I ended up making an investment in Swirlds, the company. And then, I couldn’t help myself but drop everything and join the team full-time to help take this thing from obscurity to the point where people know who we are.

Andrew Masanto: I’m Andrew Masanto. I know a lot of people here – a lot of friendly faces. Thank you for coming out. I was traditionally a lawyer. I was trained in common law at the University of Sydney in Australia. I moved to London to work for an English law firm – same law system.

While in London, I saw the craze of the Internet and started a company called Altitude Shoes – it’s still running right now. I sold that, and then got into the marketing side of things with a company called Higher Click – at the beginning of the SEO and PPC boom. I sold that, and the way I know Jordan is that he took my place at the company – he was the incoming C-level executive that basically took my place. That’s how we first met. Jordan eventually left that company and started Buffered.com, which he sold earlier this year, as you just heard.

I had enough of company building at that point, so I moved to LA to go to music school. I thought I’d just retire for the rest of my life. Then, I came to NY, and got an email from an MIT lecturer that saw that I was interested in blockchain. I had some interest in blockchain – I knew about the technology. I’m a bit of a mind geek and like reading about stuff like that. He introduced me to this concept called Hashgraph. I had a call with Mance and thought it was very interesting.

I had been investing in multiple things before, and then really went down this rabbit hole. When I got to the very bottom, I realized this technology could be really transformational. At that point in time, similar to Jordan, and many of the people who have joined us – I said, “Look, I really have to help this thing reach the light of day because it is a technology that is ripe – that is meant to happen”.

Also, Leemon Baird is truly a spectacular individual. He is one of the most impressive people I’ve ever met. He’s a very humble guy, a consummate teacher, and he invented the Hashgraph, after what I hear was years of going down rabbit holes empty handed until one day reaching down, and pulling out one of the biggest rabbits you’ve ever seen – which I think this is.

Demetri: Thank you, Andrew. Thank you, Mance and Jordan. Why don’t we first start off with an overview of what Hashgraph is?

Mance: Well, first of all – it’s not blockchain. Blockchain, as a term, refers to both a data structure – a chain of blocks of transactions, as well as a consensus algorithm – an algorithm where a community of members, who each have a copy of the chain can come to an agreement about what new block to put on top of the chain, so that the community can stay in consensus.

Hashgraph is not a chain. It’s a graph, as you might expect. The graph is created by each member of the community, contributing at any point in time, as they desire, transactions to the community. The community receives transactions all of the time. Every member, when they feel like it, submits a transaction to the community – that transaction then flows throughout the community. Each member of the community receives transactions, and they construct this graph. As new transactions flow on top of the graph and get added or stacked up on top of each other – at some point in time, very quickly, the community comes to an agreement on the order of transactions within this graph.

That sounds like magic, and looks like magic, but actually works very well. The advantage of this method is that there is no Proof-of-Work. It doesn’t have the inefficiencies that Proof-of-Work/blockchain has. It’s extremely fast because the members of the community can contribute transactions whenever they want, and the community all receives those and comes to consensus on the order very quickly. It’s not bottlenecked because there is no chain. It’s a graph, so we can achieve hundreds of thousands of transactions per second, for example, within a single peer-to-peer network. With scale, we can go to millions, or to whatever you want to achieve in scale with multiple shards or networks of communities of computers.

Demetri: You talked a little bit about blockchain there – can you go into some more detail on what blockchain is and what some of the security considerations are with respect to distributed ledger technology?

Mance: Sure. Let’s talk about databases for a moment. The assumption has always been up until 2008 that one legal entity or party has administrative control of all the databases. Amazon, for example, would have control of all of its master databases in a multi-master system that manages their store. Amazon would never think to give Google a master [database] from the Amazon cluster of databases – that is not something that ever occurred to anybody.

What blockchain is all about is doing exactly that. It’s giving masters away to other legal entities that may actually be competitive with you. It is certainly not the case that you are all going to necessarily have your interests aligned. So, if it’s the case that we’re sharing this database and each one of us has a local copy, we all have to come to an agreement on the order of the transactions and apply them in the correct order.

Well, with that new security requirement, you have to then begin to think about a few types of attacks that you wouldn’t otherwise think about.

One is that no single party in our community should be able to change the order of the transactions – the history of the order of the transactions, or, what we would normally call immutability of transactions.

Another one is that no single party in our community should be able to execute an attack that would stop the flow of transactions across the community – a distributed denial of service (DDoS) attack, which could actually bring down the network.

Finally, it should be impossible for any single party to have the influence over the order of the transactions prior to the community coming to an agreement on the order of those transactions. In other words, it shouldn’t be possible to bribe somebody to change the order of the transactions in your favor – this is what we would call fairness.

These are the fundamental concerns that arise when you move from traditional database technology all run by a single-party to this multi-party database model where we’re all sharing the data layer and we are maybe even unknown to each other.

This is what the whole blockchain space is all about.

Demetri: So, what blockchain or distributed ledger technology attempts to do is to bring disintermediation to the process of consensus – the process of determining when and how a change has occurred in the network, as well as where and how storage happens, correct?

Mance: If you think about this model, what we can do is have a shared database where we each are running the same application on our local computers. If the network and the rate at which we can process transactions and come to agreement on the order of those transactions is fast and secure enough secure enough, then we can begin to think about the “anti-cloud”, right?

What the whole cloud movement is about is having central servers that are running applications (SAS applications) that we use our computers to connect to and those central applications are providing the storage and security where we all have to trust that third party that manages it on our behalf.

What we’re talking about here [with distributed consensus] is that for the first time being able for each of us to run the same application. A good example might be World of Warcraft, or some game. We each download the distributed equivalent of World of Warcraft. We install it on our local computer. We fire it up, and a thousand of us, for example, are playing in this game with all of the transactions being written to the state of the game flowing throughout the community. The magic of the math makes it possible for the system to totally be secure in ways that allows us to not have to have a third party running a server to make sure that we’re not cheating.

We no longer have to trust a central server, much like we’ve seen recently in the rise of cryptocurrencies.  In the world of cryptocurrency, we no longer have to trust a central bank to keep the ledger. The ledger is the database that we each have a copy of. And the magic of the math ensures that no single party can change the contents of that ledger. The community has to agree on the contents of that ledger. In other words, the community agrees on the order of the transactions that we apply to the ledger so that we all have a ledger that stays in sync.

Demetri: So, let’s get into the evolution of distributed consensus from what we see now with cryptocurrencies and talk more about what makes Hashgraph unique. There’s Proof-of-Work with blockchain and other ways to drive consensus, but Hashgraph is a very different system, correct?

Jordan: With Hashgraph, we like to talk about ushering in what we feel will be the fourth generation of distributed ledger technology.

The first generation was Satoshi Nakamoto’s white paper. I suspect that most of you are familiar with Bitcoin and what that introduced to the world – distributed trust. This whole notion that we don’t know each other and we’re on opposite sides of the globe, but we can trust each other, in frankly, a trustless environment. That was really revolutionary.

Generation two was what is this thing? It’s a ledger – what else can we put on these ledgers? Perhaps, a title deed to property, stock certificate, and/or a cryptocurrency that holds value.

Generation three became these smart contracts – how do we transact fairly with each other within this distributed architecture? How do I send you my cryptocurrency and you give me the title deed to your property, and we both know that’s going to happen? This is what Ethereum is all about.

Generation four is markets and fairness. With all of this value coming on chain, we’ve got to now figure out a way to transact with each other fairly.

When I first introduce Hashgraph, I like to introduce the property of fairness, in that we’re talking about a whole new genre and world of applications that will be made possible because of this technical breakthrough – it is really immense.

Demetri: Define fairness for our audience in technical terms – what do you mean by fairness?

Jordan: When I talk about fairness, I mean specifically that we don’t necessarily trust each other, and for example, we need to know who came first in a game – if you shot me, before I shot you. That’s got to register, and we all have to agree and come to a consensus on that state of reality.

Mance: When it comes to fairness, the requirement is that there’s no single party that has the opportunity to influence the order of transactions prior to the point where the community comes to an agreement on the order of transactions.

In blockchain, for example, you have all of the miners collecting all of the transactions. They put them into a “block”, and then they compete with each other for the opportunity to publish that block before the rest of the miners on top of the chain.

The problem is that the miners unilaterally determine the order of the transactions within the block – so, they do influence the order that the community comes to agreement on with respect to the transactions. That’s the fairness argument.

Jordan is absolutely right – when it comes games, knowing which one of us actually pulled the trigger first makes a big difference. You need to be able to preserve the order of the transactions that actually flow into the network. That’s the fairness property.

Demetri: Clearly, this fairness property would be necessary in something like a stock market, monitoring traffic, or an auction site, correct?

Andrew: Fairness is necessary to do all of these types of things. You’re not going to compete in a stock market, where you don’t have mathematically guaranteed fairness. I’ve even participated in some of these ICO’s before where I’ve sent money in, and I don’t know what order the miners are putting my transaction in – I’m not in control of the fairness, and neither is the network. Again, the miner determines who comes first.

Outside of fairness, though, I just want to make the point that the real innovation for me [with Hashgraph] is speed, to be honest. I know Jordan mentioned fairness, but if you think about 1999 where we were using 56K modems – you couldn’t visualize where the Internet would go and what applications would be available as we see now with broadband fiber optic cable.

This is why I’m excited about this [Hashgraph] – the application use cases haven’t even been invented yet. Of course, I can think from my mind’s eye that there would be markets, auctions, and games because you’d have fairness and speed, but with speed, the sky is the limit. When I first discovered Hashgraph, the most exciting thing for me was speed.

Demetri: Hashgraph addresses both fairness and speed with two real innovations, correct? Using a gossip protocol to essentially gossip about gossip, and this idea of virtual voting – how does it all work?

Mance: So, how does Hashgraph work?

There is a category of consensus algorithms in the academic literature that we refer to as pure “voting-based” algorithms. They have fantastic security properties – they achieve the gold standard of security in this world of distributed systems, reaching what we call asynchronous Byzantine Fault Tolerance (aBFT).

The problem with this category of algorithms is that they are impractical in their bandwidth requirements. Conceptually, the way they work is that each peer/member of the network is asked to cast a vote on the same question. When they cast that vote, they send their vote to every other member of the network, and when every other member of the network receives a vote, they then have to send an acknowledgement that they’ve received a vote to every other member of the network. Then, there may be multiple rounds of this process that’s required in order to come to an agreement on the order of a transaction or two. In other words, the bandwidth requirements blows up and grows exponentially fast in the number of votes that are required. For this reason, they [voting-based algorithms] aren’t actually used in practice. There are no practical pure voting-based algorithm implementations at the scale that we need for what we’re talking about here.

The question then is how you achieve the same level of security with these fantastic theoretical properties [from a voting-based algorithm] without casting votes over the Internet or over the network?

The way we do it is this – we know that if [distributed] databases are going to stay in sync, then anytime there’s a transaction that needs to be written to a database, those [transactions] have to go to everybody. All transactions have to go to everybody. That’s the minimum bandwidth required, just by the laws of nature.

Leemon’s moment of inspiration – his “a-ha” moment – is what happens when you add a tiny amount of information on top of that [transaction message]. So, Alice talks to Bob and passes Bob a transaction (or update) that she’s made to her local database, and at the same time tells Bob who she spoke with last (i.e. a transaction she received from Ed, for example). Basically, when she passes a transaction to Bob, she also tells Bob “Oh, I last spoke to Ed”, and everybody [in the community] does the same thing.

It turns out you can take this information about who has spoken to whom and you can put it together in such a way that you have a picture of all of the communication, or gossip, that has taken place across the network. The picture is a picture of the gossip that has taken place across the entire network and everybody has that same picture.

We all get the same information, and we all build up the same graph. This graph is the Hashgraph. We’re passing the information along about who we spoke with last, essentially gossiping to you about who I’ve gossiped with in the past. It’s gossip about gossip – that’s the Hashgraph.

We then take this Hashgraph and we use it as an input into one of these pure voting-based algorithms. We take the algorithm and we can just run it in our head. We use the Hashgraph as an input, and we’re able to calculate, for any given point in time, what each member of the community would have voted for if they were to have cast their vote over the network. But, they don’t actually have to [vote]. Because we’re using the same Hashgraph and the same pure voting-based algorithm, we all come up with the same answer. In other words, we have consensus almost for free.

Demetri: What you’re saying is that the entire network is sharing information, and each time it shares information, it’s syncing up. Basically, I come to you and say, “This is all the stuff I know,” and you say, “This is all the stuff I know,” and then we fill each other in on the small things we don’t know. Those are the transactions, along with the metadata of who I’ve talked to and when [the hashes]. These are represented as events in memory in the Hashgraph. Each node on the network can then essentially virtually vote, eliminate the entire bandwidth constraint, and come to consensus with mathematically proven asynchronous Byzantine fault tolerance, correct?

Mance: That’s exactly what I’m saying. The important point here is that little bit of information that I said we add on top of the transactions. Technically, it is two hashes – a very tiny amount of information that can actually be compressed down even more. In other words, we achieve the bandwidth limit [in terms of speed]. I don’t know how you do better than what we do [with Hashgraph] in terms of bandwidth requirement.

And, because we’re using a pure voting-based algorithm, we achieve the gold standard in security, uniquely in the market. Hashgraph has security properties that are unmatched in the market. There is no other [consensus] algorithm in the market that is practical and achieves asynchronous byzantine fault tolerance. And, we’re fair – we’re the only algorithm in the market that has a formal definition of fairness that is mathematically proven and guaranteed.

I personally believe that Leemon has solved the problem of distributed consensus. One hundred years from now, there may be dozens of variants on the Hashgraph, but they will be on the Hashgraph. You cannot do better than asynchronous BFT, and you can’t do better than the bandwidth limit of transactions with two hashes.

Demetri: At this point, let’s go ahead and open it up to questions for our audience members.

Audience Question: With respect to fairness, can you go into more detail about how the transactions are put in order within Hashgraph?

Mance: The Hashgraph represents the gossip about gossip that’s taking place in the network. The transactions themselves actually just go along for the ride. They’re the payload in the nodes of this graph.

If you have two nodes that look like they’re at about the same vertical location in this graph, the community needs to agree which one came first. How do we do this?

We first find a community time stamp associated with each node. Essentially, if I have a transaction that I gossip out to the network, the other nodes will receive a time stamp at that received time. The entire population of the network receives my transaction and everyone in the network time stamps it.

We then find the median time stamp – not the mean or the average – but the median time stamp. Because we all are doing the same thing, we all come to the same conclusion about what that median time stamp is on the transaction. We call this the consensus time stamp.

If I lie about the time stamp and then pass that into the network, you can’t trust me. But, in assuming not more than a third of the community is evil, we can trust that the community is telling the truth about the time at which they received the transaction. If we find the median of the received time stamps, that becomes the community time stamp on the transaction. Every transaction has a community time stamp. Once you have that, you have the order. That’s the consensus order of the transactions.

Demetri: What if a node on the network gets knocked offline and then comes back online a few days later? How do you manage fairness in that scenario?

Mance: If a node drops offline for any amount of time and then comes back online, that node will simply ask the other nodes to send the latest consensus state to sync back up. Once you have the latest consensus state, then you just take the events in the Hashgraph that have occurred since that last community consensus state (which was signed by the community) and just update your local state object with latest set of transactions.

What this means is that you can actually throw away the part of the Hashgraph that’s already represented in the state object. In other words, it’s not like blockchain where you have to record the entire chain from the Genesis block on. You only have to keep the portion of the Hashgraph in memory that’s recent since the last community signed state object.

Demetri: You’re talking about short-term memory. It’s like you’re a fish and you’re able to dump that data all of the time. You don’t need to know everything that happened years ago because of the nature of the consensus. You’re not getting probabilistically closer to consensus. You know that you’re sure at a certain point in time, and you know you’ll never need to look at it again. So, you can essentially dump it is what you’re saying?

Mance: That’s right. All of the history is represented in the state. The community comes to agreement on what the state looks like at a point in time. After they’ve done that, you can get rid of all the transactions that were reflected in that state because the state implicitly already reflects all of that history.

Audience Question: So, this is very interesting and very new for me. I just wanted to ask with Hashgraph – what exactly prevents the system from getting overwhelmed by bad actors? From what I understand with Bitcoin’s security, the fact is that the system is so big, there is no bad actor(s) powerful enough to overwhelm that system. It seems like Hashgraph is more efficient and has solved the scalability problem, but what’s to prevent people from overwhelming the system, like at a poker game? If it’s me and four other people trying to beat me, I don’t stand much of a chance.

Mance: Yes, so there is a math theorem that states that it’s not possible to do better than one third of the actors in your population being evil, if they collude. If all of the evil actors collude, then up to a third of them [within the community] can be evil, and you can still come to consensus. That’s not just true for Hashgraph – that’s true in general. So yes, if you’re playing poker and there are five of you, with four being evil, then they’re going to win. That’s just going to happen. Again, this is not specific to Hashgraph in any sort of way. But, it is true that the larger the population is, the more secure the system becomes as a whole because it becomes more difficult to attract at least a third of the population being evil to collude.

Again, though, it has nothing to do whether it’s Bitcoin, EOS, Tezos, Ethereum, Hashgraph, or anybody else – if more than a third of your population is evil, then they can do bad things.

Also, this may be a little controversial, but I disagree with the premise that Bitcoin is too big to be compromised. It is possible, hypothetically, for China to shut down the great firewall and partition the system, nationalize the miners, and kill the whole chain. Technically, they could just wipe it out if they chose to do so.

Demetri: Mance, just to clarify, you’re talking about a non-permissioned network with Bitcoin, correct?

Mance: Yes, right. The market is split into two halves, correct. There are the public networks that allow anybody to download the software and participate as a full node. And, this is what we’re talking about – if not more than a third of the full nodes are evil, then you have some security. If more than a third are evil, then you’re out of luck. That’s independent of whether it’s public or private.

The question just raised is on the private side. For example, when I spoke previously about World of Warcraft, it may be that you and a dozen of your friends download the software and stand up the network. You’re all on the network and you have the security that we’re describing. You’re using your resources on your computer because you want the privilege of being able to participate in the game. So, in a permissioned network, there are other motivations for wanting to participate as a full node. In any case, there will be some governing system that decides who can participate in that network or not.

With the public networks, it’s just the opposite – it’s open. Anybody can participate that wants to.

Demetri: This brings us to a good junction in the conversation. You guys have had a lot of success on the permissioned and enterprise side. I do want you to mention a little bit of what you can about that because it’s impressive and important to note. Also, what you can tell us about a public ledger and if there are plans for that?

Mance: Sure. So, the company is about two years old and we raised our first money in Q4 of 2015. We went to work building the product. Ping Identity, a major identity player in the market was the first money into the company. It was my organization that built the identity application within Ping on top of the first version of this technology – that identity application has been proposed to solve a technical problem in the world of identity to the appropriate standards body, the Open ID Foundation.

In November of last year [2016], I moved over from Ping to Swirlds full-time and realized we needed market validation. We were a brand new technology and a brand new algorithm. Nobody had heard of us, and we were making some really strong claims. I remember over and over going in and saying to people what I’ve just said to you, and they’d say, “No way, it’s too good to be true – something’s wrong”. So, we had to have market validation.

The credit union industry saw what was happening [with blockchain] in the banking industry and decided they wanted to do the same thing. The industry wanted a platform to use for the six-thousand credit unions in North America that represent a hundred and five million members.

I decided that I wanted to compete for the business, so I approached the industry and told them what we had. They said the same thing – “No way, it’s too good to be true…but, we’ll let you compete”. So, we competed. It was Leemon and I – a couple of guys competing with IBM and Hyperledger for the business of the whole credit union industry.

We got to the first round of PoC’s, and in February of this year [2017], we won. Then, in May [2017], the industry decided to stand up a new organization called, CULedger. CULedger is a new legal entity, a credit union service organization that is now capitalized and is about to announce its first CEO. You can think of it as the equivalent of R3 for the credit union industry, as opposed to the banks. That technology is being rolled out in the new year, and it’s all built on the Hashgraph. We won. We were able to win against Goliath based on the strength of the technology – certainly, not based on the size of the company or anything of that matter. We were able to demonstrate real value in the [Hashgraph] algorithm. The platform that we built solved fundamental problems that they were not able to solve with the existing platforms in the market today.

We won that business and since then, the pipeline has just swelled on the permissioned side. We’re working with the largest banks and the largest payment processing organizations in the world. We’re addressing real permissioned use cases.

Jordan: The distinction I want to quickly make is that Swirlds is the company that created the IP. What we do is help companies use our tech stack to create private, permissioned-based networks. These organizations are standing up their own nodes, managing them, and then, they can build their vertical applications on that stack.

Demetri: What is it like to develop on the platform and how does someone get an SDK?

Jordan: That’s a really good question. We actually just went out to TechCrunch Disrupt, and the reason we were there was to actually sponsor the hackathon. This was a big “What if?” for us. We were really nervous that no one would develop on it [Hashgraph] – keep in mind, we’re a newer company coming out of stealth, and don’t have fully built out documentation like a Heroko, Wikipedia, or other companies. We just put the SDK out there, and as it turned out, in 24-hours we had some amazing applications built, like a real fully built out MMORPG (massively multiplayer online role-playing game). If you go to DEVPOST SF Disrupt 2017, you can see a bunch of these projects.

The winner we gave a prize to was a team that actually built a distributed Sotheby’s architecture where you could actually submit a bid via an SMS API called hyphenate and submit a bid to the consensus algorithm, which would pick the winner.

Accenture, who also sponsored the hackathon, actually gave their prize to a Hashgraph project, who we also wanted to pick. This prize went to a project called Ground Zero, which effectively used a mesh network type architecture to ensure that in a disaster, like a hurricane, the network would stay up and people could get an S.O.S. signal out to emergency services.

Demetri: Can you guys talk at all about a public side with Hashgraph?

Andrew: We haven’t announced a public side yet. Everyone is so interested in the public side – it is so clear that this technology has massive implications for public side. Coming back to that security question from earlier – this is the most interesting question when it comes to the public side. Obviously, with permissioned networks, you can secure the nodes. But, with the public side – that’s where it [security] really counts.

If you join our Telegram Channel and wait a couple of weeks, I guarantee you will love the answer. But, I cannot talk about that right now.

Demetri: The public side does seem to be the most exciting thing that I think everyone here is really like jazzed up about learning. Can you share anything else?

Mance: There is one thing I don’t mind addressing at this point. If you have asynchronous BFT, like in Hashgraph, this mean that you have one-hundred percent certainty (within a few seconds or less depending on various factors) on the order of transactions and you know for sure that they will never change. You have a math proof that you can take to a court of law that will stand up. You know for sure that the order of transactions will never change. It turns out that if you have aBFT within a single network (shard), you can actually scale this in ways where the level of security is maintained across the entire network.

This is not the case with the existing platforms today. In blockchain, you never come to one-hundred percent certainty that the order of transactions will never change. It’s probabilistic – with each new block that goes on top of the chain, the probability for change goes down (i.e. the coin that I’ve paid to you [the merchant] disappears).

With Hashgraph, we’re starting with the very best building block in building a scalable system which allows you to build a total system that is unparalleled in the market. If you start with the best bricks, you’re going to get the best house. Asynchronous BFT is the key to that ability to scale.

There are also a number of issues that you have to address in order to have good governance of a public network. Security is one of those. Another one is the ability to have technical controls in place that make it possible for the governing body to actually govern. If it’s the case that the governing body can’t ensure or compel the users of the technology to adopt the changes in any meaningful way, then it doesn’t matter what form of governance you have.

In Bitcoin, we have bicameral governance with the core developers and the miners. In Ethereum, we have a benevolent dictator in Vitalik [Buterin]. In both cases, we see chaos in the markets because the chains will fork or split into competing chains. Vitalik made an interesting comment earlier this summer in a podcast with a16z where he basically said the best that we can do [with Ethereum] to move the community forward is by taking advantage of each crisis as it occurs. That’s what we have today in terms of governance. There are solutions to that problem and we will talk about those in the coming weeks.

Audience Question: With Bitcoin, when you verify a transaction, it emits bitcoins onto the network. Would you be able to apply something similar with Hashgraph, where an underlying structure verifies transactions and then emits new currency onto the network?

Mance: From our perspective, what Swirlds sells is a consensus server. In the same way that every company in the world today uses an HTTP (hypertext transfer protocol) server to serve up web applications, I expect that every company in the future will also use a consensus server to serve up distributed applications.

A cryptocurrency is just an application. What we have is the ability for developers to create arbitrarily complex distributed applications using Java. There is no special scripting language like Solidity or anything else. You just use the full power of Java to build your application and the platform handles all the communications and handles all the consensus. You have business logic. You have a database that stores the state of your system (i.e. state of the world in a game, or a list of wallets and balances in a cryptocurrency). Your business logic creates transactions and passes them to the platform. The platform gossips those transactions out to the entire community. The community comes to agreement on the order of the transactions and the platform then hands the transactions in consensus order back to your application and you update your state. That’s it.

A cryptocurrency would be built on top of the platform in an obvious way. I don’t mean to minimize it, but it’s really not that complicated. It’s not. In the Bitcoin/blockchain codebase, everything is smooshed together. In the case of Hashgraph, we’ve got a nice layered architecture that separates the consensus and the communications out from the distributed application in total.

Audience Question: How secure is the gossip protocol compared to something like PGP (Pretty Good Privacy) encryption? How clean is the encryption compared to PGP?

Mance: We don’t create our own primitives. It is cryptographically secure. The nodes throughout the community are using standard crypto to ensure that the community knows that the transaction actually came from me and that it hasn’t been tampered with. It’s cryptographically secure using all the standard techniques that that the entire industry uses. There are no new crypto primitives in our implementation.

Audience Question: Most of us in here are familiar with the revenue model that the creators of Bitcoin, Ethereum, and Ripple Network had in mind when they created their products. Can you talk to us a bit about the revenue model that Swirlds has in mind as it relates to Hashgraph?

Mance: Swirlds addresses the permissioned use case – so, Swirlds sells those consensus servers to enterprise customers.

On the public network, it’s not the same. Let me let me touch on two things. When there is a public network, it will be the case that the participants in the network are all compensated for their participation. The miners today [in blockchain] participate in a lottery where the miner who solves the crypto puzzle first gets paid a bucket of coins as well as transaction fees associated with the block. In this other model [Hashgraph] that we’re talking about, all the people that participate in the network will get paid a rate. They will get paid for how much bandwidth, storage, and CPU cycles are contributed, so there will be no ambiguity. If a 16-year old kid stands up a node in their basement and runs that 24/7, they’re going to make a certain amount of money, and if all goes well, they’ll pay for their car payment with that money.

But, this is an entirely separate legal concern. This would be a different company from Swirlds. Swirlds is not going to make that much money from this public network.

Andrew: I just want to emphasize that I wouldn’t have gotten involved with this if we weren’t going to create a public utility that is basically free for everyone to use. This is the thing that I think will change the world.

There is a revenue model where Swirlds will charge customers [on the permissioned side]. It is yet to be announced on the public side, but there won’t be any licensing fees. 

To be honest, this is what interests me the most. It’s not Proof-of-Work. I love Bitcoin by the way – it is the forefather, but it is currently eating up 0.1% of the world’s electricity. It’s a massive amount of energy and a huge problem. What if it [Bitcoin] becomes even more proliferate? At a couple of percent, you’re looking at contributing to the melting of global icecaps, etc.

Bitcoin is also not a currency because it’s too slow. It’s a store of value. But, when you have something fast enough, you can become a currency at that point in time. Then, you can really help fix problems of hyperinflation or mismanaged monetary policy, which is a big passion of mine.

Mance: It’s going to look very similar to what you’re already familiar with. Application developers will build distributed applications, and when they make API calls, they will pay for those API calls with the currency. There’s no license – you just use the currency to make API calls.

Audience Question: From an investor perspective, do you plan to have an ICO?

Jordan: The answer to that is no. We’re not like anything else in the market. In fact, we’re very careful about that from the sense that what we’re trying to address are much larger market problems. So, the short answer is no, we’re not doing an ICO in the purest sense of what you see with these ERC-20 tokens. What we are doing is releasing a public protocol that you’ll be building on top of. We’ll share more information about that when we can.

And, this is just more of a commentary on what’s happening in the market, but these ICO’s are the evolution of venture capitalism, and I think it’s really important. Not all of the ICO’s are shit. There’s a lot of really great projects out there, but what we’ve got here [with Hashgraph] is not theoretical. It’s not a white paper to raise money so that we can launch a project. We already have that. Again, an ICO in the purest sense right now is to raise money or to distribute a token – we’re not using it as such.

You can go right now to Swirlds.com and download the SDK. We have six demo apps fully built that you can play with and take apart.

Audience Question: Let’s say all of your dreams come true and Hashgraph becomes the leading distributed ledger technology – can you paint us a little bit of a vision of ten years from now? What does it look like and why are things so much better?

Jordan: I want to try and do Leemon justice here. If you were to ask him, he’d tell you that one-hundred years from now, Hashgraph, or some version of it, would serve as the trust layer that sits over TCP/IP – the trust layer on top of which you’ve got a world of distributed applications that can do any number of different things, many of which we haven’t even thought up yet.

I’m going to make the claim that you’ll have multibillion dollar companies built on our tech stack because of what we enable just with property of fairness. That’s just not possible with existing DLT technology. I’m really excited for what that looks like.

I think as a company, what we will serve and introduce will be an answer to a lot of the volatility and lack of trust among the mainstream with current public chains. I want to be really careful here because I don’t want to sound like we don’t like the Bitcoin core, Vitalik, or Joe Lubin. We love those guys.

I think at the end of the day, though, we’re all trying to push the future of this new distributed web and infrastructure. Many chains may continue to exist with interoperability, but I think the dominant technology will win, and that will be Hashgraph.

Andrew: I don’t want to even claim that we’re going to be the dominant technology. If we perform the functions that I want us to perform (i.e. throughput, fairness, and security), I want to enable a whole bunch of use cases, especially the ones that I can foresee now – distributed MMO’s, auctions, and markets.

Going beyond that, I’d love to create a currency which can actually be used as a currency as opposed to an asset class. This would enable Third World countries that have poorly managed monetary policy to rely on a standard of exchange that isn’t manipulated by the government. I come from Indonesia, so I have experienced that firsthand. Again, this is something I’m very passionate about.

In terms of the rest of the world, there are countless use cases. Everything from identity, to organic farming, to supply chain. Everything that blockchain has, but when you add throughput and fairness, it opens up a whole new world of other stuff.

Mance: I think that distributed consensus technology is required for the Internet of Things (IoT) to reach its potential. Today, IoT is not grown up.

I want to get to a place where I can say which vendors in the market can bid for my business when my light bulbs burn out, for example. Well, if it’s Amazon and Wal-Mart, they need to be able to query what light bulbs have burned out and where they’re located. They should be able to get that information so that they’re able to set up a marketplace – then, I can get the best price and they automatically ship me the light bulbs.

All of that needs a distributed services directory that enables markets. It requires fairness. It requires throughput. It requires everything that we’ve talked about here tonight and it’s not possible to do it with Bitcoin and Ethereum as designed today.

Demetri: I want to wrap things up, but make a quick comment on distributed ledgers and distributed technology. We are living increasingly in a world that is intermediated by layers of technology. If we want to continue to operate as human beings in this increasingly technological future, relying on servers gives us absolutely no security in terms of how we’d like to live our lives. It’s like the classic big brother case.

Distributed systems are a way forward. Technology isn’t perfect and I think it’s a work in progress, but this is why I’m attracted to this technology. This is why I wanted to put this panel together because I think these guys may actually have solved the problem of scale, which would enable the use of distributed ledger technology to enter the mainstream. That would be a total game changer, not just in terms of the amount of money that will be saved, but also the level of disintermediation of everything in ways that will be really profound.

I want to thank our incredible panel and everyone for coming tonight.

Please check out our website, www.hiddenforcespod.com – you’ll be able to listen to my interview with Leemon Baird, where we talk more about his Hashgraph invention. I’ve also published a consolidated transcript from numerous conversations that I’ve had with him on the website.

I’m looking forward to having Leemon here next time for a follow-up event, where the community will be able to ask him questions directly.

Thank you.

See article at HiddenForcesPod.com